A principal obstacle to probative study of osteoarthritis (OA) and its treatment has been the lack of reliable imaging methods by which to noninvasively evaluate articular soft tissues over time. Using recently developed magnetic resonance imaging (MRI) sequences and MR image subtraction techniques, we have been able to generate images that, unlike conventional MRI, clearly discriminate articular cartilage, hypertrophic synovial tissue and joint effusion from each other in the articular constituents. Such a capability would permit more objective and reproducible evaluations of disease activity and treatment response in clinical trials of antiarthritic therapies, than are currently possible with available clinical, laboratory or radiographic methodology. The long-term objective of the proposed investigation is to appraise the applicability of these quantitative MRI techniques for monitoring the progression of cartilage loss in patients with OA. This will be accomplished in a longitudinal study of 25 women with OA of the knee and 10 age-and sex-matched control subjects. Women 65 years and older with simultaneous hand and knee involvement will be selected in an effort to improve disease homogeneity and increase the risk of rapid cartilage loss. Subjects will receive MRI examinations of the knee at baseline and every 9 months for 2.5 years using a high-resolution, 3D-acquired sequence. Images obtained with and without pulse magnetization transfer (MT) will be substrated and used in individual 3d reconstructions of cartilage, effusion and any synovial tissue. The volumes of each of these articular constituents will be quantified from these 3D reconstructions at each time point using a previously validated method. Standardized radiographs and clinical evaluations based on both physician-acquired data and patient self assessments will also be obtained at each point. Short-term reproducibility of the MRI technique for quantifying cartilage volume will be expressed as the standard deviation of the mean of duplicated baseline measurements in 10 of the osteoarthritic patients. Residual variation about linear regression lines fitted to the 4 repeated measurements of cartilage volume for individual patients with time as the independent variable will be used as a measure of the long-term reproducibility of the technique, incorporating errors related both to technical precision and nonlinearity of the cartilage loss. The associations between clinical, radiographic and MRI outcome variables will be described in this study would be a valuable adjunct to therapeutic trials of potential treatments of OA, and offers an intriguing new research tool with which to study the pathophysiology of this and other cartilage-destroying processes, as well as the normal physiology of cartilage. The proposed study may also provide useful information about the natural progression of cartilage changes in OA of the knee, and the importance of associated synovial hypertrophy and joint effusions to clinical examination and patient outcome.